To solve this disadvantage, this paper presents a new approach on ID3 algorithm——the information gain of attribute-value pairs——to optimize the decision tree, this alg.

To overcome the shortcomings of information loss and reduction mistakes in traditional method of rules reduction,a new algorithm of rules reduction for real value attribute information system was proposed.

The sensitivity analysis to attribute values in multiple attribute decision making with intervals and applied in submarine;

Because of many subjective and objective reasons, the determining of attribute values are often affected by uncertain, imprecise and inconstant factors, that can influence the decision-making result directly.

Improved ID3 algorithm based on attribute values;

With respect to the sensitivity analysis to attribute values in multiple attribute decision making problems, firstly, the concepts on the criticality degree of alternative and the most critical alternative were introduced.

The attribute value is the quantity or the nature of the attribute,which always is described by the quantity phrases,the nouns and the verbs.